Algorithmic Research Group

[Algorithmic Research Group @algoresearch_]

We have been building the "virtual lab" mentioned in (https://t.co/n1wGYsiNGL) and letting agents iterate on it. It will be important for measuring RSI progress and for accelerating automated safety research

[Anthropic @AnthropicAI]

Our internal data shows Claude is accelerating AI development—a possible path to recursive self-improvement, or AI autonomously building a more capable successor.

It’s happening faster than we thought, and the implications deserve greater attention. https://t.co/OVVPJO7VQx

[Algorithmic Research Group @algoresearch_]

Link: https://t.co/gGgpPZ5c1R

[Algorithmic Research Group @algoresearch_]

We've set up @AISecurityInst's Inspect platform as a lightweight remote eval service: Postgres-backed job queue, Dockerized API + worker pool, remote submission, and the existing Inspect web UI for viewing runs. It keeps the core Inspect workflow intact, but makes it practical to run many evals on a shared
host. It acts like @METR_Evals vivaria, but has Inspects awesome tooling. Link below:

[Algorithmic Research Group @algoresearch_]

🤗 Link: https://t.co/ru2holJ6Jp

[Algorithmic Research Group @algoresearch_]

We're releasing 'ai-sft' on @huggingface: a 34GB supervised fine-tuning dataset for training models on AI research tasks. 2.7M examples spanning research code generation, scientific QA, and technical problem solving, built from our research-focused data collections. Each example includes structured fields for task family, grounded context, source tracing, loss weighting, and quality flags. Train/val splits included. Link below:

[Algorithmic Research Group @algoresearch_]

whoops! push the wrong version - should be looking even better now.

[Algorithmic Research Group @algoresearch_]

We're releasing 's2orc-safety' on @huggingface: a AI safety slice of our s2orc-enriched dataset with 16,806 papers across jailbreaks, prompt injection, red teaming, model security, privacy, robustness, alignment, and more.

Each paper is enriched with structured fields for reproducibility, safety taxonomy, experimental details, practicality, normalized model/dataset/metric names,
code-link metadata, and more. Link below:

[Algorithmic Research Group @algoresearch_]

https://t.co/bdGY5B06QR

[Algorithmic Research Group @algoresearch_]

AI Safety should be open and well funded

[Matthew Kenney @baykenney]

Here are my thoughts: https://t.co/FSj3lQ0OVs

[Algorithmic Research Group @algoresearch_]

We dont have a moat now. Time to pack it in 😭

[Matthew Kenney @baykenney]

Holy shit! An intern was using claude code and made a bunch of our repos public! A thread on the fallout 🧵

[Algorithmic Research Group @algoresearch_]

AlgorithmicResearchGroup's GitHub Sponsors profile is live! You can sponsor us to support AlgorithmicResearchGroup's open source work 💖 https://t.co/SCiLc99Etn

[Algorithmic Research Group @algoresearch_]

We now support both @METR_Evals task standards and @AISafetyInst Inspect standards, split across two repos

[Algorithmic Research Group @algoresearch_]

We've added a UK AI Safety Institute (@AISafetyInst) inspect variant to make it easier to run DMLB. Link below:

[Algorithmic Research Group @algoresearch_]

These tasks are tough - we think a broad, open-ended set of research tasks mined from real world repos is the best way to measure progress in automated research. Link below:

[Algorithmic Research Group @algoresearch_]

https://t.co/dVnZw6S5KW

[Algorithmic Research Group @algoresearch_]

perfect for @karpathy style auto-research agents or large scale analysis of CS trends over time

[Algorithmic Research Group @algoresearch_]

We're releasing S2ORC CS Enriched, a dataset of 1.1 million computer science papers from Semantic Scholar's S2ORC corpus with LLM-generated enrichment fields added to every row.

The base dataset has the full paper text, abstracts, authors, references, citation counts, and venue metadata. We added structured enrichment columns on top: paper summaries, classification, methods used, results, models, datasets, metrics, limitations, and GPU compute details where reported.

The enrichment makes it possible to do things that are hard with raw paper text alone, like filtering for papers that used a specific method, or finding papers that report GPU hours, or building training data for models that need to understand the structure of research papers rather than just their text.

1,118 parquet files, 44 GB total. Available on @huggingface

https://t.co/Mp2s7suxEE

[Algorithmic Research Group @algoresearch_]

Sneak peek at our experimental interface for multiagent work

[Algorithmic Research Group @algoresearch_]

1.5k downloads in a few days. Not bad for a 54.7GB dataset 📈

[Algorithmic Research Group @algoresearch_]

1.5k downloads in a few days. Not bad for a 54.7GB dataset 📈

[Algorithmic Research Group @algoresearch_]

Contributions welcome!
https://t.co/hbuWXq2Cmh

[Algorithmic Research Group @algoresearch_]

We're excited to release HF agent - a multi-agent @huggingface model-selection and fine-tuning system that investigates current options, searches the web and Hugging Face, and produces Markdown recommendation reports with citations and code snippets. Link below

[Algorithmic Research Group @algoresearch_]

This was a fun weekend little side project to help navigate the open source repos. Uses the web and HF apis to make suggestions, provide code snippets and more. Reports are saved to @huggingface buckets, and it serves up a Spaces static site for your reports